Automated classification of stars, galaxies, and quasi-stellar objects (QSOs) is essential for large-scale astronomical surveys, where reliable classifiers must remain stable across survey releases, class imbalance, and incomplete observational inputs. In this work, we present a Brain-Inspired Quantum Machine Learning (BIQML) framework for SDSS star–galaxy–QSO classification as a step toward robust source classification for next-generation photometric pipelines. The model combines a spiking Leaky Integrate-and-Fire (LIF) controller with a classically simulated variational quantum circuit (VQC), in which the controller sequentially selects quantum gates using measurement feedback; all circuits are classically simulated, and no hardware-level quantum advantage is claimed. The framework is trained and internally tested on SDSS DR17, with SDSS DR14 used as an external cross-release test set, and is benchmarked against a broad suite of baselines including Random Forest, Extra Trees, HistGradientBoosting, Logistic Regression, Linear Support Vector Machine, and Multilayer Perceptron classifiers under a repeated ten-split protocol with paired statistical significance testing. In the full-feature setting, BIQML is statistically on par with the strongest classical baselines, reaching 97.36 ± 0.05 % accuracy (96.90% balanced) on SDSS DR17 and 98.76 ± 0.09 % accuracy (98.41% balanced) on the external SDSS DR14 test set, where the leading tree ensembles span 97.3–97.8% and 98.8–99.4%, respectively. To probe robustness when key observational channels are unavailable, we introduce a low-sensor ( g , r , i ) no-redshift setting in which spectroscopic redshift, the u - and z -bands, and metadata are removed. Under this reduced-feature condition, BIQML remains among the top-performing models on a single controlled split (0.805 balanced accuracy, versus 0.64–0.80 for the classical baselines), achieving 85.58% accuracy on SDSS DR17 and 93.44% accuracy on SDSS DR14, suggesting that it remains competitive, degrading no more than the strongest ensembles, when discriminative information is scarce. A component-level ablation isolates the contributions of the VQC, the spiking controller, and the measurement-feedback loop, while both an analytic noise screening and a finite-shot, IBM-calibrated device-noise benchmark show that BIQML predictions remain stable under representative depolarizing, amplitude-damping, readout, and realistic NISQ-style noise channels. Together, these results position BIQML as a competitive and robust quantum-inspired architecture for astronomical source classification under cross-release shift and limited observational information.
Automated classification of stars, galaxies, and quasi-stellar objects (QSOs) is essential for large-scale astronomical surveys, where reliable classifiers must remain stable across survey releases, class imbalance, and incomplete observational inputs. In this work, we present a Brain-Inspired Quantum Machine Learning (BIQML) framework for SDSS star–galaxy–QSO classification as a step toward robust source classification for next-generation photometric pipelines. The model combines a spiking Leaky Integrate-and-Fire (LIF) controller with a classically simulated variational quantum circuit (VQC), in which the controller sequentially selects quantum gates using measurement feedback; all circuits are classically simulated, and no hardware-level quantum advantage is claimed. The framework is trained and internally tested on SDSS DR17, with SDSS DR14 used as an external cross-release test set, and is benchmarked against a broad suite of baselines including Random Forest, Extra Trees, HistGradientBoosting, Logistic Regression, Linear Support Vector Machine, and Multilayer Perceptron classifiers under a repeated ten-split protocol with paired statistical significance testing. In the full-feature setting, BIQML is statistically on par with the strongest classical baselines, reaching 97.36 ± 0.05 % accuracy (96.90% balanced) on SDSS DR17 and 98.76 ± 0.09 % accuracy (98.41% balanced) on the external SDSS DR14 test set, where the leading tree ensembles span 97.3–97.8% and 98.8–99.4%, respectively. To probe robustness when key observational channels are unavailable, we introduce a low-sensor ( g , r , i ) no-redshift setting in which spectroscopic redshift, the u - and z -bands, and metadata are removed. Under this reduced-feature condition, BIQML remains among the top-performing models on a single controlled split (0.805 balanced accuracy, versus 0.64–0.80 for the classical baselines), achieving 85.58% accuracy on SDSS DR17 and 93.44% accuracy on SDSS DR14, suggesting that it remains competitive, degrading no more than the strongest ensembles, when discriminative information is scarce. A component-level ablation isolates the contributions of the VQC, the spiking controller, and the measurement-feedback loop, while both an analytic noise screening and a finite-shot, IBM-calibrated device-noise benchmark show that BIQML predictions remain stable under representative depolarizing, amplitude-damping, readout, and realistic NISQ-style noise channels. Together, these results position BIQML as a competitive and robust quantum-inspired architecture for astronomical source classification under cross-release shift and limited observational information.
Modern astronomical surveys such as the Sloan Digital Sky Survey (SDSS) produce millions of optical spectra. Classical deep learning already solves the coarse star-galaxy-quasar classification with high accuracy. This thesis asks whether quantum machine learning adds anything once the task is pushed to the hard, fine-grained subclasses where classical models begin to struggle. It insists on a fair comparison where every quantum model is matched in trainable parameters and run through an identical pipeline against a classical counterpart built with equal care. We first train a 14-million-parameter convolutional baseline that reaches 84.5% accuracy on a 62-class SDSS problem and reproduces the GaSNet-II benchmark at 92.5% on the same 13-class subset used in that study. This baseline also identifies the hardest class pairs and serves as a frozen feature extractor. We then run five experiments, namely a variational quantum circuit on the hardest binary pair, parameter-matched quantum and classical heads on the frozen extractor, a quantum support vector machine in a few-shot regime, a quanvolutional filter under increasing noise and a fully coherent quantum convolutional network (QCNN) as a negative control. Across all five the result is parity at best and no scalable quantum advantage. Two apparent advantages disappeared on closer inspection, one caused by an under-built classical baseline and one by a dead-ReLU artefact that a single change of activation removed. The only case where a quantum model did better was the QCNN, which held its accuracy when training data was scarce while the equally small classical model failed. This came from its small and rigid design helping it learn from few examples, not from any ability to represent richer functions or scale to larger problems. Because all circuits run on a noiseless statevector simulator, the most favourable setting for the variational methods studied here, the absence of an advantage is a conservative result. The thesis contributes a disciplined fair-comparison methodology and shows that apparent quantum advantages must be treated as hypotheses to be ruled out against carefully built classical controls.
Background: Breast cancer remains a leading cause of cancer-related mortality, and reliable computational decision support is increasingly viewed as a complement to expert pathological assessment rather than a replacement for it. Variational quantum classifiers (VQCs) and Quantum Support Vector Machines (QSVMs) have recently been promoted as candidate models for medical classification, yet most published comparisons rely on internal hold-out validation alone and report only a single point estimate of discrimination, omitting calibration, decision-analytic value, and explainability—three ingredients that any clinically credible model must furnish. Methods: We assembled a complete quantum–classical machine learning pipeline and evaluated it under a deliberately stringent protocol designed to expose, rather than conceal, the limitations of current Noisy Intermediate-Scale Quantum (NISQ)-era models. The analytical hypothesis was conservative and stated in advance; in light of saturated classical baselines on this benchmark, we did not anticipate a quantum advantage in raw discrimination, and we framed the study as a methodological probe rather than as a competition. Using the Wisconsin Diagnostic Breast Cancer (WDBC) dataset (n = 569) for development and an independent Wisconsin Original (WBC) cohort (n = 683) for external validation, we benchmarked five classical learners (XGBoost, LightGBM, CatBoost, RandomForest, RBF-SVM), two quantum models (an eight-qubit VQC implemented in PennyLane and a ZZ-feature-map QSVM implemented in Qiskit), and a stacked hybrid ensemble. The evaluation framework combined Optuna-driven hyperparameter optimisation, internal–external cross-validation, and external validation on the independent WBC cohort. Robustness and interpretability were then probed through circuit depth and embedding rotation ablation, depolarising noise stress tests, learning curve and feature stability analysis, decision curve analysis, and dual SHAP-based explanations covering both a direct tree-based explanation and a quantum surrogate. Reporting followed the TRIPOD + AI guideline. Results: On the internal test partition, RBF-SVM achieved the highest discrimination (AUC = 0.998), with XGBoost, LightGBM, CatBoost, the hybrid ensemble, and the VQC clustering between 0.992 and 0.996; the QSVM with a ZZ-fidelity kernel underperformed substantially (AUC = 0.727). Pairwise tests for correlated ROC curves indicated that most differences among top models were not statistically significant. On the external WBC cohort, model rankings reorganised, as RBF-SVM (AUC = 0.986, 95% CI 0.946–0.997), RandomForest (0.985, 95% CI 0.945–0.996), VQC (0.983, 95% CI 0.942–0.995), and the hybrid ensemble (0.982, 95% CI 0.941–0.995) all retained near-ceiling discrimination with extensively overlapping confidence intervals. Ablation analysis demonstrated that the choice of embedding rotation is decisive—Z-rotation embeddings collapsed VQC performance to chance levels (AUC ≈ 0.50), whereas X- and Y-rotations preserved it. Depolarising noise up to p = 0.10 had a negligible effect on the VQC, and SHAP analyses converged on worst concave points, mean concave points, and worst area as the dominant predictors across both classical and quantum models. Decision curve analysis showed positive net benefit for both classical and hybrid models across the clinically meaningful threshold range, exceeding both the treat-all and treat-none reference strategies throughout. Conclusions: In the present regime, the principal contribution of QML is not raw discrimination—modern classical learners are already at the data ceiling—but the construction of a rigorous, reproducible, externally validated, and interpretable benchmarking framework in which quantum models can be fairly compared with their classical counterparts. Because evaluation was confined to curated benchmark datasets rather than real-world clinical populations, the interpretability and net benefit findings reported here should be read as benchmark-level evidence and not as a demonstration of readiness for clinical deployment.
Data scarcity remains a critical bottleneck in applying deep learning to rare disease diagnostics, where large annotated datasets are rarely available. Quantum Machine Learning (QML) has been theoretically posited to offer superior sample efficiency; however, existing benchmarks suffer from a fundamental confound: they evaluate hybrid quantum models that leverage pre-trained classical backbones, making it impossible to attribute performance gains to the quantum circuit itself. To rigorously isolate genuine quantum inductive bias, this paper introduces a controlled “Twin Architecture” benchmarking protocol. Two structurally identical classification pipelines - a Classical Twin and a Hybrid Quantum Neural Network (HQNN) share a common, scratch-trained TinyCNN feature extractor, enforcing a strict tabula rasa experimental setup that eliminates transfer learning confounds. Both models were subjected to a systematic Data Starvation Protocol across twelve stratified subsets of a Brain Tumor MRI dataset, ranging from 100% down to 1% of available training data. At the most extreme low-data regime (5% split, n=281 training samples), the HQNN retains a classification accuracy of 59.1% (macro F1-score: 0.55), while the Classical Twin degrades to near-random performance at 26.6% (macro F1-score: 0.18). Across mid-range regimes (20%-50%), the HQNN consistently outperforms the classical baseline by a margin of 8-15 percentage points. These findings provide the first controlled, transfer-learning-free evidence that Variational Quantum Circuits (VQCs) encode a statistically meaningful inductive bias, yielding an estimated5×sample efficiency advantage over classical linear classifiers under cold-start conditions. This result positions hybrid quantum architectures as viable diagnostic tools for rare pathologies where labelled data is fundamentally limited.
Quantum neural networks (QNNs) have recently attracted remarkable attention as a potential alternative to classical and deep neural network (NN) models particularly under constraints imposed by noisy intermediate-scale quantum (NISQ) devices, which support only shallow circuits with few quantum bits (qubits) and limited variational depth. However, applying QNNs to natural language processing (NLP) remains challenging because text representations are typically high-dimensional, while NISQ classifiers can only encode a small number of features at a time, requiring classical compression to enable stable training and flexible circuit configuration. In addition, many existing studies integrate QNNs into rich-text representation pipelines, where the quantum circuit is often implemented as a variational quantum circuit (VQC). Here, the VQC acts as an additional component within a larger model, and it remains unclear how the quantum component improves the classification task. Therefore, in this work, we propose a fully reproducible and compact QNN pipeline that provides a controlled setting for studying the contribution of the VQC to text classification. The model is trained end-to-end with a lightweight classical compression layer, and the overall study is supported by systematic ablation analyses of the compression layer, the classifier heads under a shared frozen low-dimensional representation, and quantum circuit structure. Our model uses the term frequency-inverse document frequency (TF-IDF) features with linear compression followed by a shallow four-qubit, two-layer VQC, implemented in PyTorch and PennyLane for end-to-end training. It is evaluated on five benchmark text classification datasets against classical baselines, including logistic regression and multilayer perceptron (MLP) models. Experimental results show that the proposed QNN achieves a competitive performance across all evaluated datasets, yielding the best results on three of the five benchmark datasets.
Parkinson’s disease (PD) is a progressive neurodegenerative disorder affecting millions worldwide, where early and accurate diagnosis plays a critical role in improving patient outcomes. While classical machine learning methods have shown strong performance in PD classification using clinical and acoustic data, they remain limited in capturing complex nonlinear feature interactions in high-dimensional spaces. This study explores the application of quantum machine learning (QML) techniques for PD prediction, focusing on Variational Quantum Classifiers (VQC) and Quantum Support Vector Machines (QSVM), and compares their performance with strong classical baselines including Logistic Regression, SVM, Random Forest, Grid Search–optimized models, and XGBoost. Experiments were conducted using stratified train–test splits and validated through cross-validation and statistical testing. Results show that classical models achieve competitive performance, with XGBoost reaching 91.9% accuracy. However, quantum models demonstrate slightly improved performance, with the VQC using EfficientSU2 achieving the highest accuracy of 93.1%, followed by the Real Amplitudes VQC (92.6%) and QSVM (92.2%). These improvements are consistent across evaluation metrics including F1-score and recall. Statistical analysis using 10-fold cross-validation indicates that the VQC (EfficientSU2) shows a significant improvement over classical baselines ( p = 0.031), while other quantum models do not achieve statistical significance over XGBoost ( p > 0.05). McNemar’s test on the independent test set confirms no significant difference in final prediction distributions between quantum and classical models. Overall, the findings suggest that quantum machine learning models are competitive with state-of-the-art classical approaches, offering marginal but consistent performance gains. This study highlights the potential of quantum-enhanced learning frameworks for biomedical classification tasks, while also emphasizing that their advantages remain incremental rather than definitive under current simulation-based implementations.
Depression is a clinically heterogeneous disorder comprising subtypes such as melancholic, atypical, anxious, and unspecified, each characterized by distinct symptom profiles and treatment responses.Accurate identification of these subtypes is essential for precision psychiatry and optimizing therapeutic outcomes.This study investigates the potential of quantum-inspired feature representations to enhance the classification of depression subtypes from clinical data.A synthetic dataset of 5000 patients was generated, simulating realistic demographic, psychometric, behavioural, and biological variables (e.g., BDI, HRSD, cortisol, sleep, anxiety, BMI).Using standardized and quantumtransformed features, we trained and evaluated four classification models: Random Forest, Support Vector Machine (SVM), XGBoost, and a multi-layer neural network.Quantum-inspired features were derived via parameterized quantum circuits with amplitude encoding and entanglement, yielding fixedlength state vector representations.Performance was assessed using accuracy, weighted F1-score, ROC AUC, and confusion matrices on a held-out test set.Across all models, original clinical features consistently outperformed quantum-transformed features in classification accuracy and subtype separability.Statistical tests confirmed significant performance degradation with quantum features (p < 0.001).Despite this, our framework establishes a reproducible pipeline for benchmarking quantum-inspired machine learning in psychiatry.The findings highlight both the current limitations and future potential of quantum-based representations in modelling complex mental health phenotypes.This work serves as a foundation for integrating emerging quantum learning paradigms with clinically relevant, multi-class psychiatric classification tasks.